Anthropic Files Confidentially for an IPO and Turns the AI Race Into a Public-Markets Story
On June 1, Anthropic said it had confidentially submitted a draft S-1 to the U.S. Securities and Exchange Commission for a proposed IPO. The filing did not include pricing or share count, but the symbolism was obvious: one of the strongest frontier labs is now formally preparing for public markets. After its huge late-May funding round, the filing made the competition with OpenAI feel less like a private-company sprint and more like the opening of a new phase in which capital markets, index inclusion, and public scrutiny will matter almost as much as model capability.
OpenAI Brings Frontier Models and Codex to AWS
On June 1, OpenAI said its frontier models and Codex were now generally available on AWS, giving millions of AWS customers a new way to adopt OpenAI inside the security, governance, billing, and procurement systems they already use. OpenAI framed the move as a faster path from interest to implementation, especially for enterprises that want frontier models without reworking their entire cloud stack. In practical terms, it is another step in OpenAI’s effort to be bought like core infrastructure instead of an experimental AI add-on.
ChatGPT Hits 1 Billion Monthly Active App Users in Record Time
On June 2, Reuters reported that ChatGPT had crossed 1 billion global monthly active app users in May, making it the fastest app ever to hit that milestone according to Sensor Tower. The report also said Claude’s growth rate was much faster year over year even from a smaller base, underscoring how competitive the consumer AI race remains. Still, the bigger takeaway is simple: ChatGPT is no longer just a breakout product from late 2022. It is now one of the largest consumer apps in the world.
Project Glasswing Expands to 150 Organizations Across More Than 15 Countries
On June 2, Anthropic expanded Project Glasswing from roughly 50 initial partners to around 150 organizations across more than 15 countries. The company says the earlier group had already found over 10,000 high- or critical-severity flaws using Claude Mythos Preview, and the new cohort includes sectors such as power, water, healthcare, communications, and hardware. More than a product update, the expansion showed Anthropic trying to turn Mythos from a scary cybersecurity headline into a controlled, international security program.
Meta’s AI Support Bot Gets Tricked Into Handing Over High-Profile Instagram Accounts
On June 3, Reuters reported that attackers manipulated Meta’s AI support chatbot into resetting credentials and handing over access to high-profile Instagram accounts, including the dormant Obama White House page, Sephora, and a senior U.S. Space Force official. Experts told Reuters that the incident exposed a foundational design mistake: giving an AI system privileged actions without equally strong privileged controls. At a moment when tech companies are racing to automate sensitive workflows, the breach became one of the clearest warnings yet that AI support systems can also become AI attack surfaces.
Britain Forces Google to Let Publishers Opt Out of AI Search
Also on June 3, Reuters reported that Britain imposed new conduct requirements on Google Search that give publishers more control over how their content is used in the company’s AI features. Under the new rules, publishers must be able to stop their content from powering those features without losing traditional search visibility. That is a notable shift because it moves the fight over AI search from complaints and lobbying into actual regulatory obligations, and it gives publishers one of their clearest wins yet in the battle over generative summaries and traffic loss.
Washington Accelerates National-Security AI and Asks Labs for Voluntary Cyber Tests
On June 5, Reuters reported that the White House would speed the development and use of AI for national-security applications while asking leading developers to voluntarily submit their strongest models for government cybersecurity testing before public release. President Donald Trump said the administration wanted faster AI adoption across intelligence and warfighting domains, while also stressing that the technology should not be used for unlawful surveillance. The result is a distinctly American compromise: accelerate first, but try to bolt more security review onto the frontier as it moves.
